The design for this experiment was a completely randomized design with 6 replications. Unless otherwise stated, all data are presented as means and pooled S.E.M.The statistical significance of differences among the treatments (Cont., CF, AF or CAF) was evaluated using the MIXED procedure of SASs version 9.1 (SAS Institute, Inc.,Cary, NC). The statistical differences between treatments were determined by one-way ANOVA. Statements of significance were P < 0.05 unless noted otherwise.